Weights & Biases
The near-default MLOps platform, now with CoreWeave inference
Overview
An ML/AI developer platform for experiment tracking, model and dataset versioning, and LLM/agent observability via its Weave tool. Since CoreWeave's 2025 acquisition it also offers W&B Inference — a serverless, OpenAI-compatible API to run open-weight models on CoreWeave GPUs.
Pricing
Pricing shown for reference only. These figures reflect RECATOOLS research as of 13 Jul 2026 and may be out of date or incomplete. This is not financial or purchasing advice — always confirm the current price on the provider’s official website before making any decision.
- Experiment tracking
- Artifacts versioning
- Weave basics
- Usage-metered overages
- Team collaboration
- Storage $0.03/GB
- Weave ingest $0.10/MB
- Per-token inference
- Custom pricing
- SSO / governance
- Dedicated support
- Deployment options
What you can produce with Weights & Biases
- Experiment tracking and run comparison via a free, framework-agnostic SDK
- Model and dataset versioning (Artifacts)
- Weave: LLM and agent tracing / observability
- W&B Inference: serverless, OpenAI-compatible API on CoreWeave GPUs (DeepSeek, Meta, Microsoft models)
- Hyperparameter sweeps
- Shareable reports and dashboards
- Open-source client library (github.com/wandb/wandb)
ASEAN Perspective
Weights & Biases in Southeast Asia
ASEAN-region availability and pricing notes coming soon. Drop the editorial team a note via /contact/ if you can supply local context (Singapore/Malaysia/Indonesia/Thailand/Vietnam).
W&B is close to default infrastructure for ML teams — its experiment-tracking SDK is the sticky part, logging runs, metrics, and artifacts with a few lines of code. Weave extends that to LLM and agent observability with detailed traces, and since CoreWeave's roughly $1.7B 2025 acquisition there's W&B Inference: a serverless, OpenAI-compatible API to run open-weight models (DeepSeek, Meta, Microsoft) on CoreWeave GPUs without standing up your own cluster. The free tier is genuinely useful with no sales gate; Pro is $60/mo for small teams and Enterprise is custom. Watch the usage meters — storage at $0.03/GB is cheap, but Weave trace ingestion at $0.10/MB is roughly 3,000x more per byte, so heavy tracing can surprise you on the bill. Sandbox credits are limited before they become paid add-ons. If you train, fine-tune, or ship models, you'll almost certainly touch it — just model your ingestion volume before committing to heavy Weave use.
What people say
Reviewers treat W&B as near-unavoidable in machine learning. The experiment-tracking SDK is the anchor: a few lines of code log runs, metrics, hyperparameters, and artifacts, and the resulting dashboards make comparing training runs and debugging behavior straightforward. XYZEO and the official docs frame it as the AI developer platform that follows a model from datasets through experimentation to production, and its integrations with the common ML frameworks are a recurring reason teams adopt it.
Weave is the piece most newly discussed in 2026. It brings LLM and agent observability — detailed traces of prompts, tool calls, and outputs — into the same platform, aimed at teams debugging production agents rather than just training runs.
The CoreWeave acquisition reshaped the product story. CoreWeave completed its roughly $1.7B purchase in May 2025 and folded W&B into its GPU cloud with an interoperability pledge. The visible result for users is W&B Inference: log into your account, browse a model catalog, and run open-weight models — launched with DeepSeek, Meta, and Microsoft options — in the Weave Playground or through an OpenAI-compatible API, without managing your own GPUs.
Pricing drew the most scrutiny in reviews. The 2026 structure is Free / Pro ($60/mo) / Enterprise, with Enterprise effectively required for regulated orgs or teams over 50 people. Three usage meters sit underneath: storage at $0.03/GB, Weave trace ingestion at $0.10/MB, and per-token Serverless Inference. Analysts at UsagePricing flag that Weave ingestion is roughly 3,000x more expensive per byte than model storage — the single biggest cost trap for teams that trace heavily. CoreWeave Sandboxes add another wrinkle: only a limited-time credit ($10 on Free, $25 on Pro) before they become a paid add-on.
The free tier earns steady praise for being genuinely useful with no sales gate, which is a large part of why the SDK spread so widely. The consistent advice is to model your trace and storage volume before scaling up, because the metered pricing can escalate quietly once Weave is in heavy production use. For ML engineers and research teams of any size, reviewers land on W&B as a near-default choice, with cost management — not capability — as the main thing to plan around.
Summary of public user & expert reviews, compiled by RECATOOLS.
About this listing
This entry was compiled from publicly available data including Weights & Biases's official website, press releases, documentation, and reputable third-party publications. RECATOOLS is not affiliated with Weights & Biases unless explicitly stated.
Third-party AI tools update their pricing, features, availability, and policies frequently. Information here may be outdated by the time you read this — we make reasonable efforts to keep listings current, but cannot guarantee absolute accuracy.
For the latest details, please refer to Weights & Biases directly →
Spotted something out of date? Suggest an update →
More in Code & Dev Tools
